cli icon indicating copy to clipboard operation
cli copied to clipboard

Published 20 hours ago • verified publisher icon supabase

Podman support

Open probablykasper opened this issue 2 years ago • 27 comments

Feature request

Would be really nice to be able to use Podman instead of Docker. Podman is more lightweight, and from my understanding can run containers rootless

probablykasper avatar May 07 '22 19:05 probablykasper

Yes. You can! https://dev.to/zakariachahboun/supabase-self-hosting-with-podman-1mhl

zakaria-chahboun avatar Sep 23 '22 07:09 zakaria-chahboun

That's great, though personally not looking to self host, just running Supabase for local development

probablykasper avatar Sep 23 '22 07:09 probablykasper

The local development consume a lot of resources! I don't know why but it seems like there is an issue with the "authentication" part The issue executed in a loop ..

Don't forget to tell us your experience with it!

zakaria-chahboun avatar Sep 23 '22 23:09 zakaria-chahboun

I think the OP meant that when running a local dev instance using supabase start, cli can use podman to spin up the containers instead of using docker.

The main reason for this is that afaik on many Linux systems (including mine), running docker requires sudo privileges, so any command that requires interaction with the containers (e.g. supabase start, supabase db remote commit, etc.) must be ran using sudo, which can cause a variety of confusing errors such as unable to find credentials because the developer ran supabase init and supabase login as non-root. Also, I think entering the root password every five minutes is probably not a pleasant experience.

GZGavinZhao avatar Nov 28 '22 00:11 GZGavinZhao

I recently discovered that cli has the capability to use podman. It seems like under the hood, cli uses docker compose to start & orchestrate the containers, so all we have to do is to let docker compose know that we want to run our containers using podman. On Linux, the steps are similar to:

systemctl --user enable podman.socket
systemctl --user start podman.socket
systemctl --user status podman.socket
export DOCKER_HOST=unix:///run/user/$(id -u)/podman/podman.sock

Now the only problem I have is when I run supabase start, it gives me Error: unable to upgrade to tcp, received 409. This seems like a permission issue on the podman side.

Instructions for running docker compose using podman are from here: https://fedoramagazine.org/use-docker-compose-with-podman-to-orchestrate-containers-on-fedora/

GZGavinZhao avatar Nov 30 '22 00:11 GZGavinZhao

I think that the biggest issue running many docker containers on local machine is possibility to run out of memory. Podman is much more lightweight. Is there any other reason to use it instead of Docker?

proton1k avatar Nov 30 '22 12:11 proton1k

I'm getting this error running supabase start with Podman:

Seeding data supabase/seed.sql...
Error: Error response from daemon: network name supabase_network_app already used: network already exists

Running with --debug doesn't provide any useful information. Anyone has any idea? 😕

After the error above, running podman ps shows no running containers and running podman network ls only shows the default (?) network:

NETWORK ID    NAME        DRIVER
2f259bab93aa  podman      bridge

jcrqr avatar Feb 14 '23 18:02 jcrqr

The network errors are coming from here. Specifically, errdefs.IsConflict doesn't actually return true when a network with same name already exists with Podman, which raises some questions:

  • Why does this presumably work with Docker but not Podman? I'm not entirely sure, but I would guess that errdefs.IsConflict incidentally returns true when using Docker but not Podman. The errdefs does warn users:

Packages should not reference these interfaces directly, only implement them.

  • Why don't I see any networks with podman network ls? Again I'm not entirely sure, but DockerNetworkCreateIfNotExists is ran once before the error occurs with the same network ID and doesn't run into any issues. It's only when it's called the second time does it report an error. My guess is there's some tear down code that deletes the network if it fails to start, but I didn't bother looking for that code.

  • How can this be fixed? Quite easily actually. Just replace errdefs.IsConflict with a more suitable alternative, such as NetworkInspect in internal/utils/docker.go:

func DockerNetworkCreateIfNotExists(ctx context.Context, networkId string) error {
+	existing, err := Docker.NetworkInspect(
+		ctx,
+		networkId,
+		types.NetworkInspectOptions{},
+	)
+
+	// if network already exists, abort
+	if existing.ID != "" && err == nil {
+		return nil
+	}
+
	_, err = Docker.NetworkCreate(
		ctx,
		networkId,
		types.NetworkCreate{
			CheckDuplicate: true,
			Labels: map[string]string{
				"com.supabase.cli.project":   Config.ProjectId,
				"com.docker.compose.project": Config.ProjectId,
			},
		},
	)
-	// if error is network already exists, no need to propagate to user
-	if errdefs.IsConflict(err) {
-		return nil
-	}
	return err
}

Wakeful-Cloud avatar Jun 04 '23 19:06 Wakeful-Cloud

The following paragraph suggests that errdef.IsConflict is the intended use case.

To check if a particular error implements one of these interfaces, there are helper functions provided (e.g. Is<SomeError>) which can be used rather than asserting the interfaces directly.

I suspect the problem is with podman not returning the same http status code as docker daemon when there's a network name conflict. Do you mind checking with podman upstream if this is indeed the case? If so, would they be accepting PRs to address it?

sweatybridge avatar Jun 05 '23 07:06 sweatybridge

I've released a fix for the podman network issue. It's available on the beta release channel: npx supabase@beta start

Let me know if there are other incompatibilities with podman that I can help iron out.

sweatybridge avatar Jun 22 '23 07:06 sweatybridge

I tried npx supabase@beta start but got this error:

node:internal/process/promises:288
triggerUncaughtException(err, true /* fromPromise */);
^

Error: getaddrinfo EAI_AGAIN supabase_db_website
at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:dns:107:26) {
errno: -3001,
code: 'EAI_AGAIN',
syscall: 'getaddrinfo',
hostname: 'supabase_db_website'
}

Node.js v18.16.0
error running container: exit 1
Try rerunning the command with --debug to troubleshoot the error.

amr3k avatar Jul 09 '23 05:07 amr3k

I'm trying to run the Supabase CLI on Replit, which uses Nix and doesn't give the sudo access required to run Docker, so I'd also like Podman support.

@sweatybridge It looks like your PR went out in the 1.71.1 release.

The main Nix channel (23.05) only seems to have 1.62.3, but unstable has 1.75.6, so I will try that.

Nezteb avatar Jul 13 '23 03:07 Nezteb

@Nezteb sure, let me know how it goes. The nix package is maintained by the community but we are more than happy to help.

sweatybridge avatar Jul 13 '23 09:07 sweatybridge

For my particular usecase I don't think I will be able to ue Podman because I'm in a Replit unprivileged container. I will have to try running the Supabase CLI with Podman on my host laptop to confirm that your PR fixed things, but that will take me a few days. 😅

Nezteb avatar Jul 13 '23 18:07 Nezteb

I tried npx supabase@beta start but got this error:

node:internal/process/promises:288
triggerUncaughtException(err, true /* fromPromise */);
^

Error: getaddrinfo EAI_AGAIN supabase_db_website
at GetAddrInfoReqWrap.onlookup [as oncomplete] (node:dns:107:26) {
errno: -3001,
code: 'EAI_AGAIN',
syscall: 'getaddrinfo',
hostname: 'supabase_db_website'
}

Node.js v18.16.0
error running container: exit 1
Try rerunning the command with --debug to troubleshoot the error.

Got this error, too when calling DOCKER_HOST=unix:///run/user/$(id -u)/podman/podman.sock supabase start, using podman in arch. The --debug flag doesn't show any more messages.

EDIT: fixed the DNS error by installing netavark and aardvark-dns.

perroboc avatar Aug 21 '23 15:08 perroboc

@sweatybridge I am still experiencing the same issue with the beta release.

Details:

Need to install the following packages:
  [email protected]
Ok to proceed? (y) y
Supabase CLI 1.99.6

Error response from daemon: error configuring network namespace for container 8f754093eeeea5228d757190ad56f343a9de3864746277dab1cbe97672550378: CNI network "supabase_network_MyProj" not found

I then try and create the network manually:

podman network create --label com.supabase.cli.project=MyProj --label com.docker.compose.project=MyProj supabase_network_MyProj

Then trying to run again:

Supabase CLI 1.99.6
Error response from daemon: the network name supabase_network_MyProj is already used

I am on ubuntu with podman 3.4.4, which is a bit older at this point... so it is possible podman version is a variable as well.

addisonj avatar Sep 20 '23 23:09 addisonj

@addisonj I'm experiencing the same thing with 3.4.4 It seems it's been patched in recent versions.

Depending on your setup, you could to build it from source or install from Kubic's repo. The instructions for both are on podman's install page.

sameoldlab avatar Oct 06 '23 14:10 sameoldlab

There were a few more errors after this. Something like CNI bridge not working (sorry, I did not save the error) after upgrading to podman 4. Fixed by installing containernetworking-plugins.

Then there was database is not healthy when running supabase start (a more descriptive error message would be very helpful). docker logs -f supabase_db_project showed this was caused by pgsodium_root.key: Permission denied. I'm hoping there's another way to fix this but for now supabase starts with podman.service as root.

sameoldlab avatar Oct 10 '23 08:10 sameoldlab

pgsodium_root.key: Permission denied

The permission error could be due to named volumes being mounted from host to container. If you know the path to podman volume directory, try giving it read / write permission to postgres user.

sweatybridge avatar Oct 10 '23 09:10 sweatybridge

I'm not even getting any errors when I attempt to start the containers. I just get

service not healthy: [realtime-dev.supabase_realtime_sylvester supabase_pg_meta_sylvester supabase_studio_sylvester]

Debug logging shows nothing useful beyond the standard output. The container logs for those three services are dumped but don't really show anything either: realtime-dev.supabase_realtime_sylvester:

realtime-dev.supabase_realtime_sylvester container logs:
02:55:23.311 [info] == Running 20210706140551 Realtime.Repo.Migrations.CreateTenants.change/0 forward
02:55:23.315 [info] create table tenants
02:55:23.322 [info] create index tenants_external_id_index
02:55:23.328 [info] == Migrated 20210706140551 in 0.0s
02:55:23.390 [info] == Running 20220329161857 Realtime.Repo.Migrations.AddExtensionsTable.change/0 forward
02:55:23.390 [info] create table extensions
02:55:23.396 [info] create index extensions_tenant_external_id_type_index
02:55:23.401 [info] == Migrated 20220329161857 in 0.0s
02:55:23.408 [info] == Running 20220410212326 Realtime.Repo.Migrations.AddTenantMaxEps.up/0 forward
02:55:23.408 [info] alter table tenants
02:55:23.411 [info] == Migrated 20220410212326 in 0.0s
02:55:23.414 [info] == Running 20220506102948 Realtime.Repo.Migrations.RenamePollIntervalToPollIntervalMs.up/0 forward
02:55:23.420 [warning] Replica region not found, defaulting to Realtime.Repo
02:55:23.456 [debug] QUERY OK source="extensions" db=0.3ms
SELECT e0."id", e0."type", e0."settings", e0."tenant_external_id", e0."inserted_at", e0."updated_at" FROM "extensions" AS e0 WHERE (e0."type" = $1) ["postgres_cdc_rls"]
02:55:23.456 [info] == Migrated 20220506102948 in 0.0s
02:55:23.462 [info] == Running 20220527210857 Realtime.Repo.Migrations.AddExternalIdUniqIndex.change/0 forward
02:55:23.462 [info] execute "alter table tenants add constraint uniq_external_id unique (external_id)"
02:55:23.465 [info] == Migrated 20220527210857 in 0.0s
02:55:23.468 [info] == Running 20220815211129 Realtime.Repo.Migrations.NewMaxEventsPerSecondDefault.change/0 forward
02:55:23.469 [info] alter table tenants
02:55:23.471 [info] == Migrated 20220815211129 in 0.0s
02:55:23.475 [info] == Running 20220815215024 Realtime.Repo.Migrations.SetCurrentMaxEventsPerSecond.change/0 forward
02:55:23.475 [info] execute "update tenants set max_events_per_second = 1000"
02:55:23.481 [info] == Migrated 20220815215024 in 0.0s
02:55:23.490 [info] == Running 20220818141501 Realtime.Repo.Migrations.ChangeLimitsDefaults.change/0 forward
02:55:23.491 [info] alter table tenants
02:55:23.492 [info] == Migrated 20220818141501 in 0.0s
02:55:23.498 [info] == Running 20221018173709 Realtime.Repo.Migrations.AddCdcDefault.up/0 forward
02:55:23.498 [info] alter table tenants
02:55:23.499 [info] == Migrated 20221018173709 in 0.0s
02:55:23.502 [info] == Running 20221102172703 Realtime.Repo.Migrations.RenamePgType.up/0 forward
02:55:23.502 [info] execute "update extensions set type = 'postgres_cdc_rls'"
02:55:23.503 [info] == Migrated 20221102172703 in 0.0s
02:55:23.506 [info] == Running 20221223010058 Realtime.Repo.Migrations.DropTenantsUniqExternalIdIndex.change/0 forward
02:55:23.506 [info] execute "ALTER TABLE IF EXISTS tenants DROP CONSTRAINT IF EXISTS uniq_external_id"
02:55:23.508 [info] == Migrated 20221223010058 in 0.0s
02:55:23.513 [info] == Running 20230110180046 Realtime.Repo.Migrations.AddLimitsFieldsToTenants.change/0 forward
02:55:23.513 [info] alter table tenants
02:55:23.514 [info] == Migrated 20230110180046 in 0.0s
02:55:23.518 [info] == Running 20230810220907 Realtime.Repo.Migrations.AlterTenantsTableColumnsToText.change/0 forward
02:55:23.518 [info] alter table tenants
02:55:23.522 [info] == Migrated 20230810220907 in 0.0s
02:55:23.526 [info] == Running 20230810220924 Realtime.Repo.Migrations.AlterExtensionsTableColumnsToText.change/0 forward
02:55:23.526 [info] alter table extensions
02:55:23.529 [info] == Migrated 20230810220924 in 0.0s
02:55:23.532 [info] == Running 20231024094642 :"Elixir.Realtime.Repo.Migrations.Add-tenant-suspend-flag".change/0 forward
02:55:23.532 [info] alter table tenants
02:55:23.533 [info] == Migrated 20231024094642 in 0.0s
02:55:24.360 [debug] QUERY OK db=1.8ms queue=118.1ms idle=0.0ms
begin []
02:55:24.384 [debug] QUERY OK source="tenants" db=0.4ms
SELECT t0."id", t0."name", t0."external_id", t0."jwt_secret", t0."postgres_cdc_default", t0."max_concurrent_users", t0."max_events_per_second", t0."max_bytes_per_second", t0."max_channels_per_client", t0."max_joins_per_second", t0."suspend", t0."inserted_at", t0."updated_at" FROM "tenants" AS t0 WHERE (t0."external_id" = $1) ["realtime-dev"]
02:55:24.439 [debug] QUERY OK db=1.7ms
INSERT INTO "tenants" ("external_id","jwt_secret","max_bytes_per_second","max_channels_per_client","max_concurrent_users","max_events_per_second","max_joins_per_second","name","suspend","inserted_at","updated_at","id") VALUES ($1,$2,$3,$4,$5,$6,$7,$8,$9,$10,$11,$12) ["realtime-dev", "iNjicxc4+llvc9wovDvqymwfnj9teWMlyOIbJ8Fh6j2WNU8CIJ2ZgjR6MUIKqSmeDmvpsKLsZ9jgXJmQPpwL8w==", 100000, 100, 200, 100, 100, "realtime-dev", false, ~N[2023-11-27 02:55:24], ~N[2023-11-27 02:55:24], <<165, 86, 32, 78, 29, 51, 75, 35, 174, 167, 248, 212, 42, 119, 216, 160>>]
02:55:24.447 [debug] QUERY OK db=5.7ms
INSERT INTO "extensions" ("settings","tenant_external_id","type","inserted_at","updated_at","id") VALUES ($1,$2,$3,$4,$5,$6) [%{"db_host" => "f23Hm+RKKCxIK6ehAjN45GlQ0FDUt0uPXpwEKlJrfrg=", "db_name" => "sWBpZNdjggEPTQVlI52Zfw==", "db_password" => "sWBpZNdjggEPTQVlI52Zfw==", "db_port" => "+enMDFi1J/3IrrquHHwUmA==", "db_user" => "uxbEq/zz8DXVD53TOI1zmw==", "ip_version" => 4, "poll_interval_ms" => 100, "poll_max_changes" => 100, "poll_max_record_bytes" => 1048576, "publication" => "supabase_realtime", "region" => "us-east-1", "slot_name" => "supabase_realtime_replication_slot", "ssl_enforced" => false}, "realtime-dev", "postgres_cdc_rls", ~N[2023-11-27 02:55:24], ~N[2023-11-27 02:55:24], <<16, 110, 81, 172, 26, 192, 71, 186, 129, 66, 231, 218, 92, 169, 236, 206>>]
02:55:24.457 [debug] QUERY OK db=10.4ms
commit []
02:55:25.607 [info] Elixir.Realtime.SignalHandler is being initialized...
02:55:25.607 [notice] SYN[[email protected]] Adding node to scope <Elixir.Realtime.Tenants.Connect>
02:55:25.607 [notice] SYN[[email protected]] Creating tables for scope <Elixir.Realtime.Tenants.Connect>
02:55:25.608 [notice] SYN[[email protected]|registry<Elixir.Realtime.Tenants.Connect>] Discovering the cluster
02:55:25.608 [notice] SYN[[email protected]|pg<Elixir.Realtime.Tenants.Connect>] Discovering the cluster
02:55:25.608 [notice] SYN[[email protected]] Adding node to scope <users>
02:55:25.608 [notice] SYN[[email protected]] Creating tables for scope <users>
02:55:25.608 [notice] SYN[[email protected]|registry<users>] Discovering the cluster
02:55:25.608 [notice] SYN[[email protected]|pg<users>] Discovering the cluster
02:55:25.608 [notice] SYN[[email protected]] Adding node to scope <Elixir.RegionNodes>
02:55:25.608 [notice] SYN[[email protected]] Creating tables for scope <Elixir.RegionNodes>
02:55:25.608 [notice] SYN[[email protected]|registry<Elixir.RegionNodes>] Discovering the cluster
02:55:25.608 [notice] SYN[[email protected]|pg<Elixir.RegionNodes>] Discovering the cluster
02:55:25.608 [warning] Replica region not found, defaulting to Realtime.Repo
02:55:28.607 [debug] Tzdata polling for update.
02:55:28.813 [info] tzdata release in place is from a file last modified Fri, 22 Oct 2021 02:20:47 GMT. Release file on server was last modified Tue, 28 Mar 2023 20:25:39 GMT.
02:55:28.813 [debug] Tzdata downloading new data from https://data.iana.org/time-zones/tzdata-latest.tar.gz
02:55:28.907 [debug] Tzdata data downloaded. Release version 2023c.
02:55:29.447 [info] Tzdata has updated the release from 2021e to 2023c
02:55:29.447 [debug] Tzdata deleting ETS table for version 2021e
02:55:29.449 [debug] Tzdata deleting ETS table file for version 2021e
02:55:34.980 [info] Running RealtimeWeb.Endpoint with cowboy 2.10.0 at :::4000 (http)
02:55:34.988 [info] Access RealtimeWeb.Endpoint at http://realtime.fly.dev
02:55:34.989 [notice] SYN[[email protected]] Adding node to scope <Elixir.PostgresCdcStream>
02:55:34.989 [notice] SYN[[email protected]] Creating tables for scope <Elixir.PostgresCdcStream>
02:55:34.989 [notice] SYN[[email protected]|registry<Elixir.PostgresCdcStream>] Discovering the cluster
02:55:34.989 [notice] SYN[[email protected]|pg<Elixir.PostgresCdcStream>] Discovering the cluster
02:55:34.990 [notice] SYN[[email protected]] Adding node to scope <Elixir.Extensions.PostgresCdcRls>
02:55:34.990 [notice] SYN[[email protected]] Creating tables for scope <Elixir.Extensions.PostgresCdcRls>
02:55:34.990 [notice] SYN[[email protected]|registry<Elixir.Extensions.PostgresCdcRls>] Discovering the cluster
02:55:34.990 [notice] SYN[[email protected]|pg<Elixir.Extensions.PostgresCdcRls>] Discovering the cluster

supabase_pg_meta_sylvester:

supabase_pg_meta_sylvester container logs:

> @supabase/[email protected] start
> node dist/server/server.js

(node:21) ExperimentalWarning: Importing JSON modules is an experimental feature. This feature could change at any time
(Use `node --trace-warnings ...` to show where the warning was created)
{"level":"info","time":"2023-11-27T02:55:24.680Z","pid":21,"hostname":"3a9950ac0e89","msg":"Server listening at http://0.0.0.0:8080"}
{"level":"info","time":"2023-11-27T02:55:24.688Z","pid":21,"hostname":"3a9950ac0e89","msg":"Server listening at http://0.0.0.0:8081"}

supabsae_studio_sylvester:

supabase_studio_sylvester container logs:
  ▲ Next.js 13.5.3
  - Local:        http://localhost:3000
  - Network:      http://0.0.0.0:3000

 ✓ Ready in 523ms

This is with Podman 4 and after installing containernetworking-plugins

Like other users above, after the "service not healthy" message podman ps shows no running Supabase containers and podman network ls only shows the default bridge

RedChops avatar Nov 27 '23 03:11 RedChops

Adding my notes to the conversation from trying to get supabase working with podman on macOS 14 ...

  1. I had to remove the host.docker.internal:host-gateway option from supabase CLI because podman doesn't support this option. Instead they've opted to automatically add a host.docker.internal entry to every container's /etc/hosts by default. See https://github.com/containers/podman/issues/10878.
diff --git a/internal/db/start/start.go b/internal/db/start/start.go
index dd7558f..ced03d5 100644
--- a/internal/db/start/start.go
+++ b/internal/db/start/start.go
@@ -95,7 +95,6 @@ func NewHostConfig() container.HostConfig {
                        utils.DbId + ":/var/lib/postgresql/data",
                        utils.ConfigId + ":/etc/postgresql-custom",
                },
-               ExtraHosts: []string{"host.docker.internal:host-gateway"},
        })
        return hostConfig
 }
diff --git a/internal/functions/serve/serve.go b/internal/functions/serve/serve.go
index dc8fb66..66199ab 100644
--- a/internal/functions/serve/serve.go
+++ b/internal/functions/serve/serve.go
@@ -163,7 +163,6 @@ EOF
                },
                start.WithSyslogConfig(container.HostConfig{
                        Binds:      binds,
-                       ExtraHosts: []string{"host.docker.internal:host-gateway"},
                }),
                network.NetworkingConfig{
                        EndpointsConfig: map[string]*network.EndpointSettings{
  1. Now I'm getting Error response from daemon: lsetxattr /Users/wryfi/src/github.com/wryfi/supaflut/supabase/functions: operation not supported. It appears that one of the containers is trying to set an extended attribute – which macOS doesn't support but podman doesn't prevent – on a volume-mounted folder. But I can't readily tell which container is responsible and I hit my time box for investigating further.

Hope to see the supabase CLI fully functioning with podman soon!

wryfi avatar Nov 29 '23 16:11 wryfi

Also: for anyone just looking for a docker desktop alternative, colima is working for me with supabase with the docker runtime that's available in macports/homebrew.

wryfi avatar Nov 29 '23 17:11 wryfi

@wryfi Using your diff, on Linux I was able to start the local containers using supabase start --ignore-health-check. Even though podman ps shows that realtime is unhappy:

> podman ps
CONTAINER ID  IMAGE                                            COMMAND               CREATED             STATUS                         PORTS                    NAMES
9f4fdb0f28ac  public.ecr.aws/supabase/postgres:15.1.0.117      postgres -c confi...  About a minute ago  Up About a minute (healthy)    0.0.0.0:54322->5432/tcp  supabase_db_buoj
fd3d49376f2f  public.ecr.aws/supabase/kong:2.8.1                                     About a minute ago  Up About a minute (healthy)    0.0.0.0:54321->8000/tcp  supabase_kong_buoj
414830bb3ffc  public.ecr.aws/supabase/gotrue:v2.99.0           gotrue                About a minute ago  Up About a minute (healthy)                             supabase_auth_buoj
4d5a45a185b8  public.ecr.aws/supabase/inbucket:3.0.3           -logjson              About a minute ago  Up About a minute (healthy)    0.0.0.0:54324->9000/tcp  supabase_inbucket_buoj
2d059c4c9cbc  public.ecr.aws/supabase/realtime:v2.25.35        /bin/sh -c /app/b...  About a minute ago  Up About a minute (unhealthy)                           realtime-dev.supabase_realtime_buoj
c2389193d94b  public.ecr.aws/supabase/postgrest:v11.2.2        /bin/postgrest        About a minute ago  Up About a minute                                       supabase_rest_buoj
b387e3033dd3  public.ecr.aws/supabase/storage-api:v0.43.11     node dist/server....  About a minute ago  Up About a minute (healthy)                             supabase_storage_buoj
14c75c57f04d  public.ecr.aws/supabase/imgproxy:v3.8.0          imgproxy              About a minute ago  Up About a minute (healthy)                             storage_imgproxy_buoj
8b3f6dac3aa2  public.ecr.aws/supabase/edge-runtime:v1.23.0                           About a minute ago  Up About a minute                                       supabase_edge_runtime_buoj
f347e5c9df83  public.ecr.aws/supabase/postgres-meta:v0.75.0    npm run start         About a minute ago  Up About a minute (unhealthy)                           supabase_pg_meta_buoj
848ac49c7b0e  public.ecr.aws/supabase/studio:20231123-64a766a  node apps/studio/...  About a minute ago  Up About a minute (unhealthy)  0.0.0.0:54323->3000/tcp  supabase_studio_buoj

I wrote some quick tests to test realtime and realtime is still working. I inspected the logs using podman logs realtime-dev.supabase_realtime_buoj and I am not seeing anything suspicious.

I think there may be some problems with the healthchecks. Will report back when I encounter any issues. My podman installation has netavark, aardvark-dns, and cni-plugins installed as well, no sure if that would make a difference.

GZGavinZhao avatar Nov 29 '23 20:11 GZGavinZhao

Also @wryfi your issue with lsexattr looks like containers/podman#13631. You may want to check relevant commits at the bottom of that issue to see if your version of podman contains the relevant fix.

GZGavinZhao avatar Nov 29 '23 20:11 GZGavinZhao

As of today, I was still unable to run supabase env on my M2 Mac. What I did:

# install podman, initialize the machine
sudo podman-mac-helper install
podman machine stop
podman machine set --rootful
podman machine start
export DOCKER_HOST="unix:///var/run/docker.sock"
supabase start

but I end up with (and gave up):

failed to start docker container: Error response from daemon: failed to create new hosts file: unable to replace "host-gateway" of host entry "host.docker.internal:host-gateway": host containers internal IP address is empty

hnykda avatar Apr 01 '24 07:04 hnykda

It looks to be working on my Linux machine with rootless podman running as my user, but only when I ignore the health checks. When I start with health checks - it starts, I can browse around in the web UI, then it disappears. Maybe the healthchecks are incorrectly being reported as not healthy, and then the command decides to shut down the containers, and ignoring the health checks makes it work?

n3oney avatar Apr 02 '24 18:04 n3oney

Maybe the healthchecks are incorrectly being reported as not healthy, and then the command decides to shut down the containers, and ignoring the health checks makes it work?

That is possible. Is there any logs you can share when start fails due to health check?

sweatybridge avatar Apr 03 '24 03:04 sweatybridge